Background and Importance Therapeutic drug monitoring (TDM) is the clinical practice of measuring drugs to maintain a constant concentration in the patient‘s blood, thereby individualising dosage regimens. TDM is mainly used to monitor drugs with a narrow therapeutic range, drugs with high pharmacokinetic variability, and drugs with a high incidence of adverse effects.

The narrow therapeutic window of lithium (0.6 – 0.8 mmol/L) requires accurate monitoring of its serum concentrations to achieve a safe and effective therapy.

Aim and Objectives To compare two pharmacokinetic programmes and analyse the precision and accuracy of lithium serum concentration adjustment.

Material and Methods Retrospective observational study including admitted patients with at least one determination of serum lithium concentration between January and December 2020 at a secondary hospital.

Electronic medical records were used to obtain the following data: lithium dosage, serum lithium concentrations, date of blood analysis, serum creatinine, renal function (calculated using the Cockcroft-Gault equation), date of birth, sex and weight.

Serum lithium concentrations were estimated using two pharmacokinetic software programs: MwPharm++ and PKS.

Accuracy and precision were evaluated using Sheiner and Beal’s prediction error theory. Accuracy was estimated with the mean prediction error (MPE) and precision with the mean absolute prediction error (MAPE) and the square root of the root mean square prediction error (RMSE). These results are accompanied by 95% confidence intervals.

The statistical significance was determined using a t-student test for comparing means.

Results A total of 79 plasma lithium levels from 18 patients were analysed, 55.6% were male, with a median age of 52.4 years [IQI: 41.7–55.4], and a median weight of 70.5 kg [IQI: 66.8–82.15]. Three patients (16.7%) had a creatinine clearance less than 60 ml/min, and 17 (94.4%) had multiple serum lithium determinations. The median number of determinations per patient were 3 IQI [2–4.5].

The following results were obtained:

Accuracy: MPE 0.02 (-0.025–0.065) and -0.02 (-0.064–0.024) for MwPharm++ and PKS, respectively.

Precision: MAPE 0.14 (0.11–0.18) and 0.12 (0.08–0.16), and RMSE 0.20 and 0.20 for MwPharm++ and PKS, respectively.

No statistically significant differences were found for MPE (p=0.22) or MAPE (p=0.40).

Conclusion and Relevance MwPharm++ and PKS showed satisfactory predictive capabilities, with no significant statistical differences. Both programs seem to be valid options, but larger studies are needed for confirmation.

Conflict of Interest No conflict of interest.